Skip to content

README.md

Latest commit

 

History

History
134 lines (92 loc) · 4.05 KB

README.md

File metadata and controls

134 lines (92 loc) · 4.05 KB

amazon-textract-haystack

PyPI - Version PyPI - Python Version

Overview

A Haystack integration for AWS Textract that extracts text and structured data from documents using OCR.

The AmazonTextractConverter component converts images and single-page PDFs into Haystack Document objects using the AWS Textract synchronous API.

Supported file formats: JPEG, PNG, TIFF, BMP, and single-page PDF (up to 10 MB).

Installation

pip install amazon-textract-haystack

Usage

Basic text extraction

Extract plain text from a document using DetectDocumentText:

from haystack_integrations.components.converters.amazon_textract import AmazonTextractConverter

converter = AmazonTextractConverter()
results = converter.run(sources=["document.png"])
documents = results["documents"]

print(documents[0].content)

Table and form analysis

Use AnalyzeDocument to detect tables and forms by setting feature_types:

converter = AmazonTextractConverter(feature_types=["TABLES", "FORMS"])
results = converter.run(sources=["invoice.png"])

documents = results["documents"]
raw_responses = results["raw_textract_response"]

Valid feature_types values: "TABLES", "FORMS", "SIGNATURES", "LAYOUT".

Natural-language queries

Ask questions about a document and get extracted answers. The QUERIES feature type is enabled automatically when you pass the queries parameter at runtime:

converter = AmazonTextractConverter()
results = converter.run(
    sources=["medical_form.png"],
    queries=["What is the patient name?", "What is the date of birth?"],
)

documents = results["documents"]
raw_responses = results["raw_textract_response"]

Queries can be combined with feature_types for both structural and question-based extraction:

converter = AmazonTextractConverter(feature_types=["TABLES", "FORMS"])
results = converter.run(
    sources=["invoice.png"],
    queries=["What is the total amount due?"],
)

In a Haystack pipeline

from haystack import Pipeline
from haystack.components.preprocessors import DocumentCleaner
from haystack_integrations.components.converters.amazon_textract import AmazonTextractConverter

pipeline = Pipeline()
pipeline.add_component("converter", AmazonTextractConverter())
pipeline.add_component("cleaner", DocumentCleaner())
pipeline.connect("converter.documents", "cleaner.documents")

result = pipeline.run({"converter": {"sources": ["scan.png"]}})

AWS Credentials

The component uses the standard boto3 credential chain. You can configure credentials in any of these ways:

  1. Environment variables (default): Set AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, and AWS_DEFAULT_REGION.
  2. AWS credentials file: Configure via ~/.aws/credentials and ~/.aws/config.
  3. IAM role: When running on AWS infrastructure (EC2, Lambda, ECS).
  4. Explicit parameters:
from haystack.utils import Secret

converter = AmazonTextractConverter(
    aws_access_key_id=Secret.from_env_var("MY_AWS_KEY"),
    aws_secret_access_key=Secret.from_env_var("MY_AWS_SECRET"),
    aws_region_name=Secret.from_token("us-east-1"),
)

Running Tests

Unit tests (no AWS credentials needed):

cd integrations/amazon_textract
hatch run test:unit

Integration tests (require AWS credentials and a test image at tests/test_files/sample_text.png):

export AWS_ACCESS_KEY_ID=...
export AWS_SECRET_ACCESS_KEY=...
export AWS_DEFAULT_REGION=us-east-1
hatch run test:integration

Contributing

Refer to the general Contribution Guidelines.